|
|
Post by knucklehead on Mar 5, 2017 20:44:37 GMT -5
I had a Rotel RMB-1085 for a while. Rotel uses the B&O IcePower modules in that particular amp. I never noticed an accentuation of the highs. FWIW anyone using a class d amp with a processor with room correction should not notice extended highs since the room correction software will tamp down out of range frequencies. I used the Rotel with a Sherbourn 7030 which has no room correction. I never felt the need to use the internal equalizer to tame the high frequencies. I've also had several Crown XLS amps over the years - same results.
|
|
|
|
Post by Casey Leedom on Mar 5, 2017 21:19:52 GMT -5
Yes, but the reason I mentioned doing an FFT comparison is that I seriously doubt that there is any kind of big non-linearity in the Frequency Response of a well-implemented Class D Amplifier based on the B&O ICEpower modules versus a classic Linear Class A/B Amplifier. I can only imagine that it's some kind of phase issue which can sometimes manifest as perceived high frequencies?
Casey
|
|
|
|
Post by Gary Cook on Mar 6, 2017 1:31:26 GMT -5
So Gary, I presume you've also listened to ICEpower Amplifiers and felt they also "accentuated the high frequencies"? It would be interesting to see what an FFT would show compared to a linear amplifier. Casey Hi Casey, I have heard a couple of B&O digital amplifiers (from memory they use IcePower modules) and they sounded OK. Admittedly in a showroom, a good acoustic show room, and I wan't super impressed, they were just OK. I never felt the need to take one home and try it in my system, although it was offered. They were playing through a set of Paradigm Signature series speakers, the shop owner knows that I like them as my uncle has a set that he purchased there. So he set it up deliberately for my benefit, even used a couple of my test music tracks (he knows me too well), and was a bit disappointed that I declined. In summary, they were OK sounding but not anywhere near good value for money as they sure as hell aren't cheap here in Australia. Cheers Gary |
|
|
|
Post by Gary Cook on Mar 6, 2017 1:41:22 GMT -5
In regards to evenness of frequency response, from my listening experience I don't believe that it's as simple as they are accentuating the volume of the higher frequencies. It's like the Oppo Sabre DAC implementations, I'm sure we are all absolutely certain that they measure as flat as technically possible to measure. But plenty of people notice (some like, some dislike) their etchiness (to use Keith's terminology), for example to me cymbals and harps etc just don't sound "right". Obviously Class D amplifiers haven't the same affected sound, but I bet they measure flat, so any amount (or type) of room correction isn't going to make diddly squat difference.
There is no substitute for actual listening and no two different rooms, speakers and ears are going to achieve the same result.
Cheers
Gary
|
|
|
|
Post by Casey Leedom on Mar 6, 2017 11:50:35 GMT -5
I was actually thinking of the experiment where one records the output of a system with a test track by mic'ing the speakers with one amplifier or another and then does a comparison of the data gathered to see what differences there are. I.e. hold everything constant except the amplifiers. Obviously one would want a really high quality microphone and ADC, etc., etc., but it would be interesting to see if we could construct a non-subjective measurement scheme for issues like this.
Casey
|
|
|
|
Post by Gary Cook on Mar 6, 2017 15:17:45 GMT -5
I was actually thinking of the experiment where one records the output of a system with a test track by mic'ing the speakers with one amplifier or another and then does a comparison of the data gathered to see what differences there are. I.e. hold everything constant except the amplifiers. Obviously one would want a really high quality microphone and ADC, etc., etc., but it would be interesting to see if we could construct a non-subjective measurement scheme for issues like this. Casey That's pretty much one of the testing processes for power amplifiers and pre amplifiers, feed a known signal in (volume and frequency) and compare it to the output. Emotiva often publish the results of that testing with the product information. With continuous improvements in testing, more and more test processes have been developed to try and make objective what our ears hear. But my personal view is that there's plenty more subtleties in audio that there simply aren't testing process and/or equipment yet invented to measure them. One of the tests that I often carry out when evaluating new gear, especially power amplifiers, is to connect one channel to the new amp and leave the other channel connected to the original amp. Switch to mono and listen, it's amazing what subtle differences can be picked up with such a simple test. Internal noise generated in the amp, introduced noise from external sources, earth loop, etc. Much more conclusive than relying on memory, playing one amp then unplugging it and replacing it with the test amp. The same test process works for comparing interconnects and speaker cables. With a pair of monoblocks, comparing power cords is also ever so easy. Strangely it's not one of the tests that I see or read about, surely professional reviewers would have thought about it. After all it's very simple, adds nothing to review time other than the actual listening and it reveals in real time subtle differences that memory may miss. Cheers Gary |
|
|
|
Post by bolle on Mar 9, 2017 8:14:16 GMT -5
I had a Sherbourn PA 7-350 and changed to Pascal Audio S-Pro 2, which are an audible upgrade for me. Also tried Hypex UCD and Icepower, both were a downgrade compared to the Sherbourn.
|
|
|
|
Post by Casey Leedom on Mar 9, 2017 11:15:55 GMT -5
So I think that the Hypex UcD was the predecessor to their current Ncore technology, right?
Casey
|
|
Lonnie
Emo Staff 
admin
Pay no attention to the man behind the curtain
Posts: 6,999
|
Post by Lonnie on Mar 9, 2017 11:29:34 GMT -5
The extended highs that a lot people hear when using class D amplifiers is all to often the interaction of the recombinant filter reacting to the varying resistance, capacitance and back EMF (or more simply put, "reactance") of the speakers. When an amplifier is tested, they are alway run on a non-inductive load bank which gives consistent, repeatable test results. But speakers are not non-inductive loads, they are reactive loads. Their impedance varies as does their capacitance and inductance. So virtually all Class D amplifiers have a recombinant filter on the output to take the PWM signals and turn them back into sine waves. These filters are for all practical purposes a speaker crossover that is turned to a 4 or 8 ohm load and set to roughly 40Khz. As with all Inductor/capactor based crossovers, their actual crossover frequency and level they pass is dependent on the reactive load attached. So in the case of a class D and a speaker, you have two crossovers run in series and both are ultimately tied to a reactive load. So what may measure flat as a board on a fixed load, will change dramatically when connected to a reactive load. The way to correct for this is through feed-back to the modulator (which is what we do in our Class D amps). Class D amplifiers by nature of how they work, do not require any feedback for stability but if you add it back into the modulator, you can get frequency error correction and as a result, they sound like a linear amp. Just my .02  Lonnie |
|
|
|
Post by geebo on Mar 9, 2017 12:20:15 GMT -5
The extended highs that a lot people hear when using class D amplifiers is all to often the interaction of the recombinant filter reacting to the varying resistance, capacitance and back EMF (or more simply put, "reactance") of the speakers. When an amplifier is tested, they are alway run on a non-inductive load bank which gives consistent, repeatable test results. But speakers are not non-inductive loads, they are reactive loads. Their impedance varies as does their capacitance and inductance. So virtually all Class D amplifiers have a recombinant filter on the output to take the PWM signals and turn them back into sine waves. These filters are for all practical purposes a speaker crossover that is turned to a 4 or 8 ohm load and set to roughly 40Khz. As with all Inductor/capactor based crossovers, their actual crossover frequency and level they pass is dependent on the reactive load attached. So in the case of a class D and a speaker, you have two crossovers run in series and both are ultimately tied to a reactive load. So what may measure flat as a board on a fixed load, will change dramatically when connected to a reactive load. The way to correct for this is through feed-back to the modulator (which is what we do in our Class D amps). Class D amplifiers by nature of how they work, do not require any feedback for stability but if you add it back into the modulator, you can get frequency error correction and as a result, they sound like a linear amp. Just my .02 Lonnie ^ Yeah, what he said. |
|
|
|
Post by mgbpuff on Mar 9, 2017 13:15:25 GMT -5
Making a type D amp sound as good as a competent Class A or A/B amp is like taking an alligator, trussing him up, covering him with hair, and teaching him to bark like a dog. Just get a dog!
|
|
|
|
Post by Casey Leedom on Mar 9, 2017 13:28:25 GMT -5
Thanks for the input Lonnie . So has anyone at Emotiva had a chance to compare some of the Class D Amplifier designs like the Bang and Olufsen ICEpower, the Hypex Ncore, this new "Pascal" design from D-Sonic, etc.? For instance, I've seen articles written by Bruno Putzeys of Hypex/Ncore fame where he extols the virtues of Negative Feedback in Amplifier Design[1,2]. Is this part of the Feedback which you talk about? And how does all of this compare with the design of the current XPA Gen3 amplifiers? As I understand it they're Class A/B being fed by a Switch-Mode Power Supply? So they're "linear" in the blades I think, but still require a Low-Pass Filter to remove high-frequency Power Supply noise (though I think the Class A/B design of the blades may do that naturally). Casey [1] Negative feedback in audio amplifiers: Why there is no such thing as too much[2] Negative feedback in audio amplifiers: Why there is no such thing as too much (Part 2) |
|
nrde
Minor Hero
Posts: 62 
|
Post by nrde on Mar 9, 2017 15:33:43 GMT -5
Making a type D amp sound as good as a competent Class A or A/B amp is like taking an alligator, trussing him up, covering him with hair, and teaching him to bark like a dog. Just get a dog! Seem to work well in studio monitors though. |
|
|
|
Post by mgbpuff on Mar 9, 2017 15:53:23 GMT -5
Making a type D amp sound as good as a competent Class A or A/B amp is like taking an alligator, trussing him up, covering him with hair, and teaching him to bark like a dog. Just get a dog! Seem to work well in studio monitors though. I guess that's O.K. if you want to monitor a studio! |
|
|
|
Post by Gary Cook on Mar 9, 2017 16:29:22 GMT -5
-------- recombinant filter ------------- Lonnie Nice one Lonnie, increasing ones vocabulary, I've not seen "recombinant" used in relation to electronics, is that alternate terminology for an L-C low-pass filter? Cheers Gary |
|
|
|
Post by bolle on Mar 9, 2017 17:08:37 GMT -5
this new "Pascal" design from D-Sonic, etc.? Pascal Audio is a company from Denmark. It is their design, D-Sonic is just one of their customers... |
|
|
|
Post by Axis on Mar 9, 2017 17:27:20 GMT -5
I like digital amps. The Onkyo TX-LR552 that is in my profile sig is no longer in service but served me for 12 years. It was a very good Digital amp that sounded great. I am using a Sony 1070 A/B amplifier of the same power rating and not feeling it like I did with that old digital amp. I think I am going to get an RMC-1 with a Mercedes Benz knob !  Emersa is old news. |
|
klinemj
Emo VIPs 
Official Emofest Scribe
Posts: 15,090
|
Post by klinemj on Mar 9, 2017 18:10:26 GMT -5
You've tried the ncores? What did you think of them? I did post up my thoughts some time ago, I recall it was round 2013 when I first experience a pair of Ncores and I have had a couple of similer experiences since. Just quickly, one example, I long time friend of mine also a professional roady back in the day, is a dab hand with a soldering iron and he bought an Ncore kit pair to DIY. Once completed he ran into some difficulties, mostly digital interference as well as distortion from one channel and an overall unimpressive sound quality. My opinion not his, he rather liked the sound, but the other issues were driving him nuts. I hopped over to his place a couple of times and with Hypex help (via very expensive phone calls) we were finally able to rectify the issues. A couple of component swaps (fixed the distortion issue) and re packaging pretty much fixed the digital interference issues (not 100%, but tolerable). He has some, what I would call, fairly mellow speakers in a pair of vintage Tannoys and plays a lot of vinyl, so the high frequency accentuation is not so noticeable. But even in that environemnt it still annoys me and I find it tiring for long term listening as well. I have tried one of his Ncores in my system, with one channel Ncore amplification and the other XPA-5, switch to mono and listen. Needless to say I didn't leave it in my system for very long. He currently works on movies as a rigger and as a result is often overseas for long periods of time. When he is around we exchange venues for our long time friends' get togethers and it's split about 50/50 who likes and who doesn't like the Ncores. So I'm not alone, the tube fanboy friend of ours (yes, everybody has one of those) in particular hates them with a passion. It's kinda like Sabre DACs, some people like them and some people don't. Cheers Gary If you heard nCores that had an issue as you describe, I can expect you would not like them. I probably would not either. Too bad you guys had to make calls to sort it out. I found the advice of users at audiocircle + email help from Hypex more than sufficient to do a good build the first time. But, it is well known that doing certain things or not doing them on a home build can make or break the result. And, it is not really hard if one follows the basics. I also do not like the Oppo sound for 2 channel. But, I do like the sound of my LH Labs DAC which also uses ess chips like oppo but different. And, the nCore sound when built properly? Excellent. Long story short...the devil is in the details. (And I love tube sound also...) Mark |
|
|
|
Post by Casey Leedom on Mar 9, 2017 19:29:55 GMT -5
|
|
|
|
Post by Casey Leedom on Mar 9, 2017 20:19:15 GMT -5
|
|
Emersa is old news.
